To address the urgent demand for autonomous rapid initial alignment of vehicular inertial navigation systems in complex battlefield environments, this study overcomes the technical limitations of traditional stationary base alignment methods by proposing a robust moving-base autonomous alignment approach based on multi-source information fusion. First, a federal Kalman filter-based multi-sensor fusion architecture is established to effectively integrate odometer, laser Doppler velocimeter, and SINS data, resolving the challenge of autonomous navigation parameter calculation under GNSS-denied conditions. Second, a dual-mode fault diagnosis and isolation mechanism is developed to enable rapid identification of sensor failures and system reconfiguration. Finally, an environmentally adaptive dynamic alignment strategy is proposed, which intelligently selects optimal alignment modes by real-time evaluation of motion characteristics and environmental disturbances, significantly enhancing system adaptability in complex operational scenarios. The experimental results show that the method proposed in this paper can effectively improve the accuracy of vehicle-mounted alignment in motion, achieve accurate identification, effective isolation, and reconstruction of random incidental faults, and improve the adaptability and robustness of the system. This research provides an innovative solution for the rapid deployment of special-purpose vehicles in GNSS-denied environments, while its fault-tolerant mechanisms and adaptive strategies offer critical insights for engineering applications of next-generation intelligent navigation systems.

In order to effectively solve the problems of ambient temperature variation significantly affects the measurement accuracy of the gravimeter and the customary temperature control scheme severely restricts the rapid start-up capability of gravity measurement, a two-layer temperature compensation algorithm (TLTCA) for a zero-length spring relative gravimeter is proposed. The upper layer temperature compensation model is constructed based on influence mechanisms and correlation analysis, and the population initialization and position update formulas of the sparrow search algorithm (SSA) used to optimize model coefficients are improved to form ISSA, which significantly improves the search capability of the global optimal solution. A large-scale measured dataset was created to train the lower layer temperature compensation model using a BP neural network (BPNN). The TLTCA integrates the advantages of the upper and lower models, improving the accuracy and stability of the gravimeter output. The temperature compensation effects of SSA, ISSA, BPNN, and TLTCA are examined under four experimental environments, namely, constant external temperature, natural temperature variation, 1 °C/min temperature rise and fall, and rapid temperature variation. The results are compared to the least squares method (LSM), commonly used in engineering practice. The experimental results demonstrate that the TLTCA has the best effect, with the standard deviation after compensation up to 0.89 μGal, 2.76 μGal, 8.66 μGal, and 12.02 μGal, respectively, which is comparable to that of the high-precision temperature control scheme and has important reference significance for enhancing gravimeter rapid start-up and environmental adaptability.

To address the issue that dynamic gravity anomaly measurement is overly dependent on GNSS and cannot be measured autonomously at this stage, this paper proposes an autonomous ground vehicle dynamic gravity anomaly measurement method based on a strapdown inertial navigation system (SINS), odometer (OD), barometer and platform gravimeter. The SINS/OD/barometer integrated navigation solution delivers high-precision navigation parameters, completes the calculation of correction terms, and performs the autonomous dynamic gravity anomaly measurement combined with the primary measurement results of the platform gravimeter. Numerical calculations provide the requirements for the application of the proposed method, and the cut-off frequency for extracting gravity anomalies is 0.02 Hz, as determined by power spectral density analysis. In order to further improve the measurement accuracy and account for dynamic errors caused by vehicle maneuvering, a long-short-term memory (LSTM) model of recurrent neural network is introduced. A series of experiments under multiple circumstances with repeated lines were conducted in Tianjin, China, and the static measurements along the line were taken using CG-5 to provide true values of gravity anomalies. The results demonstrate that the autonomous measurement scheme can achieve accuracy comparable to GNSS-assisted, and that dynamic error compensation algorithm based on LSTM improves the dynamic gravity measurements accuracy significantly without sacrificing the spatial resolution of gravity anomalies.

Vertical vibration directly impacts the measurement accuracy of a zero-length spring relative gravimeter. Vibration isolation has limited suppression effect in complicated vibration situations, and an effective and reliable vibration compensation method is crucial. The mechanism by which the vibration environment affects the gravimeter is first analyzed in this letter. In addition, an improved alternating direction multiplier method (ADMM) based on the inexact criterion and golden ratio proximity is proposed to find the optimal model coefficients of the time-delay gain simplified model of the vibration transfer function and compared with the sparrow search algorithm (SSA). The simulation and experimental results show that the method proposed in this letter has the best vibration compensation effect, and the standard deviation (STD) and mean absolute error (MAE) can reach up to 0.1424 and 0.0965 mGal, respectively, which is of great significance for improving the environmental adaptability of gravity measurement.